A simulation study published in JMIR Formative Research offers valuable guidance for researchers evaluating prevention effectiveness in respiratory and other infectious disease studies. The article, “Methods to Adjust for Confounding in Test-Negative Design Effectiveness Studies: Simulation Study,” is authored by Westat’s Elizabeth Rowley, DrPH, Patrick Mitchell, ScD, and Duck-Hye Yang, PhD, and explores how different statistical methods can reduce bias in these estimates.
The study compares multivariable logistic regression and disease risk score (DRS) approaches using simulated data. The goal: to identify which methods most accurately adjust for confounding factors that can distort real-world effectiveness findings.
The results can help researchers choose the right tools for complex study designs, ultimately improving data quality. They show that while both methods can be effective, multivariable models generally offer better coverage and precision. DRS-based models also performed well, especially when stratified, but underestimated uncertainty in some cases.
